Complexity, Generalizations of Monadic Functions, and Code Reuse

[DavidHarrison]

• Add comments on the complexity of more functions
• Reduce the complexity of findLastIndex
• Generalize replicateM and filterM to Applicative
• Reuse existing functions where reasonable in definitions

[paf31]

Is this definitely still O(n)? It only is if the singletons are concatenated to the left of the result.

[DavidHarrison]

I believe it should be the same as the above functions using foldMap, since foldMap is right fold, and concatenation to the singleton should be constant time. It does feel like a fair amount of wrapping/unwrapping, perhaps it would be better to define mapMaybe (and use it for filter as well)?

[paf31]

Looks great! I added a couple of comments.

[garyb]

I'm pretty sure this will be a lot slower than the original definition - reconstructing the entire list and taking an element vs reading it once and taking the last element?

[DavidHarrison]

I hadn't been thinking of that. I'll try using a fold.

[garyb]

Currently TCO will optimise this, a fold will be slower too.

[paf31]

Sorry this has gone unmerged for so long. Would you be interested in rebasing this on top of master? Or maybe this could be broken up into a few smaller PRs.

[DavidHarrison]

Thanks! I'll split it up into smaller PRs.

[paf31]

Ok thanks! I'll close this one then.